Photolithography or microlithography apparatus are widely used in the fabrication of microelectronic semiconductor devices and other microdevices. In photolithography, an optical system directs light energy to record a pattern at high resolution and with precise registration onto a photosensitive layer formed on a silicon wafer or other substrate. Continuing improvements in miniaturization place increasingly more challenging demands on the performance and accuracy of the optical system used for this function. Microlithography optical systems are fairly large and complex, containing a number of optical elements. A stacked annuli lens assembly arrangement is preferred for this type of optical apparatus, as described, for example, in U.S. Pat. No. 5,428,482 entitled “Decoupled Mount for Optical Element and Stacked Annuli Assembly” to Bruning et al.
Lenses of very high quality are used for microlithography. Typically, these lenses (known as “stepper lenses”) comprise a number of elements, where each lens element is very accurately mounted in a cylindrical shaped “cell” made of stainless steel. Each of these cells is fabricated to extremely tight tolerances. Mating surfaces, for example, are ground flat and parallel, so that when the lens is assembled each successive cell is bolted to the face of the adjacent cell with little or no adjustment possible. Once all the cells have been assembled, the entire lens is tested and any unwanted aberrations or image defects are discovered.
In practice, after a lens is completely assembled for the first time, it is often determined through rigorous testing that one or more of the elements must be moved slightly in the X or Y direction in order to correct the measured optical defect. This type of adjustment must be accomplished without adversely affecting the position of nearby components and without changing the position of the lens element along the optical axis. In some cases, this type of correction entails disassembly of the lens assembly, re-adjustment of lens position, re-assembly, and re-testing. As is well known to those skilled in optical fabrication, this can be a costly and time-consuming procedure subject to human error.
An alternate strategy that accommodates the need to make X-Y centering adjustments relates to design of the lens cell itself, with an inner ring connected to an outer mount. This approach is used, for example, in the complex optical mount disclosed in U.S. Pat. No. 6,191,898 entitled “Optical Imaging Device, Particularly an Objective, with at Least One Optical Element” to Trunz et al. The outer mount in this type of design supports the structure and mounts to adjacent cells in the lens assembly and the inner ring carries the lens element to be adjusted. One or more opposing adjustment screws are then used to urge the inner ring to a preferred position within the X-Y plane that is orthogonal to the optical axis (Z axis).
Although solutions using an inner ring supported within an outer mount can alleviate the need to disassemble the lens assembly when adjustment within the X-Y plane is required, there are drawbacks to this type of approach, in practice. Conventional solutions of this type can be subject to frictional forces and surface slippage during adjustment, which can contribute to undesirable and unpredictable parasitic motion, so that adjustments that are made in order to shift the position of the inner ring along one direction result in unwanted motion relative to the orthogonal direction. The amount of unwanted motion can be difficult to predict from one adjustment to the next and depends on numerous factors such as the surface contour and finish and relative rotational position of the actuator screw or other actuator shaft for both the driven and the unmoved actuator, the angle of contact between the driven and non-driven actuators and the inner ring, and the beginning and ending positions in the X-Y plane.
Thus, when using a conventional arrangement of actuators for adjusting X-Y plane positioning, results may not be satisfactory. Unwanted effects of frictional forces and variations in surface geometry at the mechanical interfaces can cause some amount of parasitic motion upon adjustment that is difficult to predict. A number of the conventional solutions proposed for X-Y plane adjustment are fairly complex and include a large number of components, increasing the risk of introducing unwanted parasitic motion when adjustments are made.